During the 9 years from 1963 to 1971, 5, 296 times of gastric cytology and 1, 972 times of endoscopic biopsy were carried out at National Cancer Center Hospital. Techniques empolied for the collection of cells were gastric lavage, abrasive balloon, endoscopic lavage, imprint smears of biopsy or surgical materials. Results of cytology and biopsy were reported by 5 classes and classes III to V were calculated as positive. In malignant cases, its diagnosis was confirmed by surgery or autopsy or made by biopsy and the patients had died without autopsy.However, in to two thirds of benign cases their diagnosis was based only on biopsy or clinical follow-up by radiology and endoscopy.
Positive rate for malignant cases was 71.1% in lavage method, 75.0% in imprint smear and 87.3% in biopsy. Falsepositive reports were encountered in 2.8%, 5.7% and 0.1% of benign cases respectively. Very low rate of fals-epositives is the most charact eristic feature of biopsy.
The positive rate in early cancers was 56.4% in lavage methad, 74.8% in imprint smear and 92.9% in biopsy.The lavage method, which may collect cells from all parts of gastric lining, is an ideal way for cytologic diagnosis of gastric cancer.However its accuracy decreases considerably when the size of a cancerous lesion becomes smaller than 3 cm.By radiology or endoscopy, detection of a focus smaller than 3 cm is not so difficult.Accordingly lavage cytology for screening of gastric malignancy was discarded gradually.In imprint smear and biopsy, on the contrary, small size of a lesion scarcely affect their accuracy.Somewhat unfavourable results in imprint smears and biopsy for cancers located at the proximal part of the stomach were caused by technical difficulty of endoscopy.However, this difficulty had been eliminated recently by the improvement of instruments.For cancer of this part, positive rate was relatively high in lavage method. Comparatively high rate of false-positives in imprint smear was considered to have been resulted partly by our poor knowledge about fresh cells excised by biopsy.Imprint smear cytology was performed as an adjuvant to biopsy.Both of them were positive in 76.8%, only imprint cytology in 3.7% and both were negative in 7.6% of malignant cases.In cytology, positive rate was somewhat low in the signet-ring cell type cancer and in biopsy it was slightly low in the tubulary adenocarcinoma.
In malignant lymphoma the result of cytology and biopsy was lower than that in carcinoma, and in leiomyosarcoma both were negative except one case of positive biopsy.In cases of borderline atypical epithelium, lavage method was usually negative, but imprint smear was positive in 33.3% and biopsy in 3.3%.In biopsy, 90.0% of them were reported as borderline cases.
Much criticism was centered to the Papanicolou's classification for expression of cytologic diagnosis. We have adoptted 5 classes to indicate our confidence about malignant neature of cells on a specimen.By lavage method, for example, the percentage of malig nant cases among those which were reported as class I was 8.4%, 19.5% among class II, 67.9% among class III, 88.4% among class IV and 99.2% among class V. We believe that a report by 5 classes is a suitable way to inform an impression of a cytologist to the clinicians concerned.
Usefullness of cytology or biopsy for diagnosis of gastric cancer was investigated by a comparison of their results to clinical diagnosis when the examination was indicated.The number of malignant case which were detected among those who were diagnosed as bengin by barium meal and/or endoscopy were 24 by lavage method, 34 by imprint smear and 50 by biopsy during these 9 years.In conclusion, superiority of biopsy for the diagnosis of gastric malignancy was unquestionable.However cytologic examination by imprint smear is indispensable as an adjuvant to it.
抄録全体を表示